With advances in genetic engineering, a large number of protein drugs have been prepared and utilized. However, protein drugs are susceptible to denaturator or proteolytic degradation in the body, and it is difficult to sustain in vivo concentrations and titers for a long period of time. Developing a technique to maintain the in vivo concentrations of protein drugs at suitable levels by enhancing protein stability is important to promote the efficacy of therapy, to help the patients who need to take their protein drug in frequent injections and to reduce the cost of treatment.
Many attempts have been made to enhance the in vivo stability of protein drugs for a long time, and such attempts include the changing a protein formulation, fusing a protein to another protein, or chemically or biologically attaching a suitable polymer to the surface of a protein.
One of such technique is making a fusion protein with the immunoglobulin Fc fragment.
The Fc fragment mediates effector functions such as complement-dependent cytotoxicity (CDC) or antibody-dependent cell-mediated cytotoxicity (ADCC), as well as antigen binding capacity that is the unique function of immunoglobulins. Also, the Fc fragment contains a sequence participating in the binding to the neonatal Fc receptor (FcRn), which plays a role in regulating serum IgG levels by increasing the transport of maternal IgG to neonates and the half-life of the IgG (Ghetie and Ward, Immunology Today 18: 592-598, 1997), and the sequence regulates the interaction between protein A and protein G. Through the fusion of this Fc fragment with a therapeutic protein, many studies have been performed to enhance the stability of the therapeutic protein.
For example, Korean Pat. No. 249572 discloses a fusion protein which is prepared by linking an IgG1 heavy chain constant region (Fc) at an amino terminal end thereof to a carboxyl terminal end of a protein, such as IL4 receptor, IL7 receptor, G-CSF receptor or EPO receptor, and producing the resulting fusion protein in mammalian cells. U.S. Pat. No. 5,605,690 describes a fusion protein comprising tumor necrosis factor receptor fused at its carboxyl terminal end to a human IgG1 Fc derivative, the fusion protein being produced in animal cells. Also, Tanox Inc. reported in U.S. Pat. Nos. 5,723,125 and 5,908,626 that a hybrid molecule comprises human interferon alpha or beta linked at its carboxyl terminal end to native human IgG4 Fc through a peptide linker and is produced in animal cells. Lexigen Inc., as described in International PCT Application Publication No. WO 00/69913, prepared a native IgG1 Fc linked at its carboxyl terminal end to the amino terminal end of human interferon by genetic recombination without the use of a linker and produced the fusion protein in animal cells. U.S. Pat. Publication No. 20030082679 discloses a fusion protein with an extended serum half-life, which comprises human G-CSF linked at its carboxyl terminal end to the amino terminal end of IgG1 Fc and is produced in animal cells. U.S. Pat. Publication No. 20010053539, U.S. Pat. No. 6,030,613, International PCT Application Publication Nos. WO 99/02709 and WO 01/03737 and European Pat. No. 0464533B1 disclose an Fc fusion protein with a longer serum half-life than a native protein, which comprises an IgG1 Fc or Fc derivative linked at its amino terminal end through a peptide linker or without a peptide linker to the carboxyl terminal end of human EPO, TPO, human growth hormone or human interferon beta, the Fc fusion protein being produced in animal cells.
These Fc fusion proteins, as described above, increase the serum half-life of a target protein, but are problematic in terms of mediating effector functions by the Fc fragment (U.S. Pat. No. 5,349,053). Through the effector functions of the Fc fragment, they fix complements or bind to cells expressing FcγRs, leading to lysis of specific cells, and induce the production and secretion of several cytokines inducing inflammation, leading to unwanted inflammation. Also, the fusion creates a new amino acid sequence at a connection region between the Fc fragment and the protein partner, which could potentially induce immune responses upon administration for a long time.
In this regard, many efforts have been made to prepare an immunoglobulin or immunoglobulin fragment having a long serum half-life but being deficient in effector functions. Cole et al. reported that, when amino acid residues of the CH2 region at positions 234, 235 and 237, known to play an important role in the binding to Fc receptors, are replaced with alanine to produce an Fc derivative having a reduced binding affinity to Fc receptors, the ADCC activity is inhibited (Cole et al., J. Immunol. 159: 3613-3621, 1997). However, in all of these variants, Fc may have increased immunogenicity or antigenicity compared to the native human Fc fragment due to the presence of unsuitable amino acids and may lose preferable Fc functions.
Among methods of deleting or reducing undesirable effector functions while maintaining high serum concentrations of an immunoglobulin, one is based on removing sugar moieties from the immunoglobulin. As described in U.S. Pat. No. 5,585,097, an aglycosylated antibody derivative as an anti-CD3 antibody can be prepared by replacing a glycosylated residue of antibodies, the asparagine residue at position 297 of the CH2 domain, with another amino acid. This aglycosylated antibody derivative exhibits reduced effector functions, but still retains its binding affinity to FcRn receptor with no change of serum half-life. However, this derivative is also problematic in terms of being potentially recognized as a foreign material and rejected by the immune system due to the production of a novel recombinant construct having an abnormal sequence. U.S. Pat. Publication No. 20030073164 discloses a method of producing an Fc derivative using E. coli devoid of glycosylation ability so as to prepare a therapeutic antibody deficient in effector functions.
The American company Amgen Inc. described, in U.S. Pat. No. 6,660,843 and U.S. Pat. Publication Nos. 20040044188 and 20040053845, a human IgG1 Fc derivative having amino acid deletions at the first five amino acid residues of the hinge region, which is fused to the amino or carboxyl terminal end of a therapeutic protein or a therapeutic protein mimicked by a peptide, and the production thereof using an E. coli host. However, this fusion protein not having a signal sequence is expressed as inclusion bodies and thus must be subjected to an additional refolding process. This protein refolding process reduces yields, and especially in a protein present as a homodimer or a heterodimer, remarkably reduces dimer production. Also, when a protein not having a signal sequence is expressed in E. coli, a methionine residue is added to the N-terminus of the expression product due to the feature of the protein expression system of E. coli. The aforementioned expression products of Amgen Inc. have an N-terminal methionine residue, which may induce immune responses upon repeated or excessive administration to the body. Also, since these fusion molecules are expressed in a fusion protein form in E. coli through linking a gene encoding a therapeutic protein to an Fc gene, they are difficult to express in E. coli, or a therapeutic protein is difficult to produce in E. coli if its expression in E. coli in a fused form results in a significant decrease or loss or activity. Further, since the fusion of two molecules creates a non-naturally occurring abnormal amino acid sequence at the connection region between two proteins, the fusion protein could be potentially recognized as “non-self” by the immune system and thus induce immune responses.
To solve these problems, the present inventors previously prepared an Fc fragment and a protein drug as separate polypeptides not using a fusion method based on genetic recombination but using the best expression systems and covalently linking the two polypeptides together to use the Fc fragment as a drug carrier. In this case, it is possible to prepare a conjugate of a glycosylated polypeptide drug and an aglycosylated Fc, which does not induce undesirable immune responses but has satisfactory properties of physiological drug activity, in vivo duration and stability.
In the above case, since it is preferable that the Fc is in an aglycosylated form, a prokaryotic expression system such as E. coli is used. Protein production methods using an E. coli expression system have several advantages compared to methods using animal cells, as follows. An E. coli expression vector can be easily constructed, thus allowing rapid evaluation for protein expression. Due to its rapid growth rate, E. coli allows mass production of a protein of interest at low cost. Also, a relatively simple expression process can be used. Thus, E. coli is more useful for commercial production than other host cells.
However, there is no report of an industrially available useful method for mass production in E. coli of immunoglobulin constant regions that are present as inclusion bodies upon overexpression in E. coli. 
Thus, leading to the present invention, the intensive and through research into a method capable of mass-producing immunoglobulin constant regions such as an immunoglobulin Fc fragment, conducted by the present inventors, resulted in the finding that, when a nucleotide sequence encoding an immunoglobulin constant region such as an immunoglobulin Fc fragment is expressed in E. coli in a form fused to an E. coli signal sequence, the immunoglobulin constant region is expressed not as inclusion bodies but in a water-soluble form in E. coli. 